I ran short benchmarks before and after KPTI and new firmware. The performance on small-size file copy, pipe throughput and system calls seem to have been significantly affected.
Before:
BYTE UNIX Benchmarks (Version 5.1.3)
System: latt: GNU/Linux
OS: GNU/Linux -- 4.14.6-1-ARTIX -- #1 SMP PREEMPT Sun Dec 17 18:51:09 UTC 2017
Machine: x86_64 (unknown)
Language: en_US.utf8 (charmap="ANSI_X3.4-1968", collate="ANSI_X3.4-1968")
CPU 0: Intel(R) Core(TM) i5-7200U CPU @ 2.50GHz (5426.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
CPU 1: Intel(R) Core(TM) i5-7200U CPU @ 2.50GHz (5426.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
CPU 2: Intel(R) Core(TM) i5-7200U CPU @ 2.50GHz (5426.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
CPU 3: Intel(R) Core(TM) i5-7200U CPU @ 2.50GHz (5426.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
20:29:20 up 53 min, 4 users, load average: 0.24, 0.36, 0.20; runlevel
------------------------------------------------------------------------
Benchmark Run: Thu Jan 04 2018 20:29:20 - 20:57:37
4 CPUs in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 37237440.1 lps (10.0 s, 7 samples)
Double-Precision Whetstone 4346.2 MWIPS (11.2 s, 7 samples)
Execl Throughput 6930.6 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 1206772.9 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 365865.8 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 2781821.8 KBps (30.0 s, 2 samples)
Pipe Throughput 2633111.8 lps (10.0 s, 7 samples)
Pipe-based Context Switching 256969.5 lps (10.0 s, 7 samples)
Process Creation 16859.3 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 9802.4 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 2590.3 lpm (60.0 s, 2 samples)
System Call Overhead 3493960.4 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 37237440.1 3190.9
Double-Precision Whetstone 55.0 4346.2 790.2
Execl Throughput 43.0 6930.6 1611.8
File Copy 1024 bufsize 2000 maxblocks 3960.0 1206772.9 3047.4
File Copy 256 bufsize 500 maxblocks 1655.0 365865.8 2210.7
File Copy 4096 bufsize 8000 maxblocks 5800.0 2781821.8 4796.2
Pipe Throughput 12440.0 2633111.8 2116.6
Pipe-based Context Switching 4000.0 256969.5 642.4
Process Creation 126.0 16859.3 1338.0
Shell Scripts (1 concurrent) 42.4 9802.4 2311.9
Shell Scripts (8 concurrent) 6.0 2590.3 4317.2
System Call Overhead 15000.0 3493960.4 2329.3
========
System Benchmarks Index Score 2051.4
After:
BYTE UNIX Benchmarks (Version 5.1.3)
System: latt: GNU/Linux
OS: GNU/Linux -- 4.14.12-1-ARTIX -- #1 SMP PREEMPT Sat Jan 6 08:04:05 UTC 2018
Machine: x86_64 (unknown)
Language: en_US.utf8 (charmap="UTF-8", collate="UTF-8")
CPU 0: Intel(R) Core(TM) i5-7200U CPU @ 2.50GHz (5426.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
CPU 1: Intel(R) Core(TM) i5-7200U CPU @ 2.50GHz (5426.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
CPU 2: Intel(R) Core(TM) i5-7200U CPU @ 2.50GHz (5426.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
CPU 3: Intel(R) Core(TM) i5-7200U CPU @ 2.50GHz (5426.0 bogomips)
Hyper-Threading, x86-64, MMX, Physical Address Ext, SYSENTER/SYSEXIT, SYSCALL/SYSRET, Intel virtualization
10:11:57 up 1 min, 1 user, load average: 0.17, 0.08, 0.03; runlevel
------------------------------------------------------------------------
Benchmark Run: Sun Jan 07 2018 10:11:57 - 10:40:15
4 CPUs in system; running 1 parallel copy of tests
Dhrystone 2 using register variables 37309489.5 lps (10.0 s, 7 samples)
Double-Precision Whetstone 4353.9 MWIPS (11.2 s, 7 samples)
Execl Throughput 6024.0 lps (30.0 s, 2 samples)
File Copy 1024 bufsize 2000 maxblocks 824374.2 KBps (30.0 s, 2 samples)
File Copy 256 bufsize 500 maxblocks 231631.3 KBps (30.0 s, 2 samples)
File Copy 4096 bufsize 8000 maxblocks 2358767.1 KBps (30.0 s, 2 samples)
Pipe Throughput 1296364.4 lps (10.0 s, 7 samples)
Pipe-based Context Switching 223604.9 lps (10.0 s, 7 samples)
Process Creation 14871.0 lps (30.0 s, 2 samples)
Shell Scripts (1 concurrent) 5008.4 lpm (60.0 s, 2 samples)
Shell Scripts (8 concurrent) 1547.2 lpm (60.0 s, 2 samples)
System Call Overhead 829213.5 lps (10.0 s, 7 samples)
System Benchmarks Index Values BASELINE RESULT INDEX
Dhrystone 2 using register variables 116700.0 37309489.5 3197.0
Double-Precision Whetstone 55.0 4353.9 791.6
Execl Throughput 43.0 6024.0 1400.9
File Copy 1024 bufsize 2000 maxblocks 3960.0 824374.2 2081.8
File Copy 256 bufsize 500 maxblocks 1655.0 231631.3 1399.6
File Copy 4096 bufsize 8000 maxblocks 5800.0 2358767.1 4066.8
Pipe Throughput 12440.0 1296364.4 1042.1
Pipe-based Context Switching 4000.0 223604.9 559.0
Process Creation 126.0 14871.0 1180.2
Shell Scripts (1 concurrent) 42.4 5008.4 1181.2
Shell Scripts (8 concurrent) 6.0 1547.2 2578.6
System Call Overhead 15000.0 829213.5 552.8
========
System Benchmarks Index Score 1382.2